SEP 9, 2026 · PREPRINT
FlowCPO: A Unified Divergence View of Preference Alignment for Flow Models
arXiv
This is an unrefereed arXiv preprint introducing a machine learning method for preference alignment in generative models, with theoretical analysis and empirical results that have not undergone peer review.
Reported
GenEval score (in-domain)0.84
OCR score (in-domain)0.87
GenEval score FlowDPO baseline0.81